Main content area

Prognostication of energy use and environmental impacts for recycle system of municipal solid waste management

Nabavi-Pelesaraei, Ashkan, Bayat, Reza, Hosseinzadeh-Bandbafha, Homa, Afrasyabi, Hadi, Berrada, Asmae
Journal of cleaner production 2017 v.154 pp. 602-613
algorithms, databases, emissions, energy flow, environmental impact, environmental indicators, life cycle assessment, municipal solid waste, natural resources, neural networks, prediction, recycled materials, recycling, topology, toxicity, transportation, waste disposal, Iran
The aim of this study is to perform a life cycle and energy flow assessments of Municipal Solid Waste (MSW). These latter are modeled using Artificial Neural Network (ANN) for integrated waste management system in Tehran Municipality, Iran. The initial data used in this study have been obtained from Waste Management Organization of Tehran Municipality, while the required data related to the background system were extracted from the EcoInvent 2.2 database. Based on the obtained results from the energy cycle analysis, energy consumptions for transportation and processing plus recycling were calculated as 227.02 and 155.95 Gigajoules (GJ), respectively for 8500 t MSW. This energy consumption has led to the production of 18,884.03 GJ energy output (recycled materials) with an energy ratio of 121.56. A high energy ratio demonstrated that recycling of municipal solid is a new gateway against energy challenges. By optimizing energy consumption, especially energy related to transportation, this ratio could increase more than before. The results obtained from Life Cycle Assessment (LCA) indicate that transportation is considered the most important hot spot in the recycling of MSW while recycling papers is the main contributor to the redaction of environmental indicators in the recycling system. More intensity is observed for indicators related to toxicity due to the lack of waste disposal and to the entry of undesirable substances in natural resources. The 5-7-7-11 structure is the best topology to predict the amount of recycled materials and the impacts of environmental emissions from MSW recycling process developed by feed-forward back-propagation ANN models which are based on Levenberg-Marquardt (LM) training algorithm. Coefficient of determination values for train are in the ranges of 0.926–0.978 and have excellent performance in predicting all the output parameters based on the input parameters.